Multi level odds ratio - help

[meatduck]

Hi,

I'm a newbie when it comes to SPSS and Statistics, and in need of some help.

I would like to know how to calculate a multi level factor odds ratio in
SPSS, for an outcome (0 vs 1 : no disease vs disease) according to X protein
serum levels (normal, borderline, high). And also if there's anyway to
compare these odds ratio to another Y and Z protein.

Thanks in advance,
Carlos



--
Sent from: http://spssx-discussion.1045642.n5.nabble.com/

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

[meatduck]

Hi,

I'm a complete newbie when it comes to this, and in need of some help.

I would like to know how to calculate odds ratio for an outcome (0 vs 1 - no
disease vs disease), of different serum X protein levels (normal,
borderline, high) in SPSS. And is there any way to compare these odds ratio
to another Y protein or Z, with the same levels (normal, borderline, high)?

Thanks in advance,
Carlos



--
Sent from: http://spssx-discussion.1045642.n5.nabble.com/

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

[Maguin, Eugene]

Second question: If you have access to data for the second protein and the methods and variables of the second study are judged to be same/similar enough to the data for the first protein, why not combine the data sets and analyze both. But if not, and you have the standard error of the coefficients from the second study and the level 1 model variable set is the same as for your data model,  why couldn't you construct a 95% or 99% confidence interval for corresponding variables in each study and see if they overlap.

First question: since you're using a multilevel model, you must have some variables to predict variance in the level 1 intercepts. If you are interested in the odds ratios for the level 1 coefficients, why wouldn't you just exponentiate the coefficients, that same as you would do for a logistic regression?

Gene Maguin


-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of meatduck
Sent: Thursday, June 18, 2020 8:05 AM
To: [hidden email]
Subject: Multi level odds ratio - help

Hi,

I'm a newbie when it comes to SPSS and Statistics, and in need of some help.

I would like to know how to calculate a multi level factor odds ratio in SPSS, for an outcome (0 vs 1 : no disease vs disease) according to X protein serum levels (normal, borderline, high). And also if there's anyway to compare these odds ratio to another Y and Z protein.

Thanks in advance,
Carlos



--
Sent from: http://spssx-discussion.1045642.n5.nabble.com/

=====================
To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

[Bruce Weaver]

Hello Carlos.  You've had one reply from Gene Maguin, who took your use of
"multi level factor" to mean that you are estimating a multilevel logit
model.  I understood it differently:  I understood that you simply meant the
explanatory variable of interest, serum level, has more than two levels.
You suggested normal, borderline, and high.  For a categorical variable with
3 levels, one of them is designated as the reference category, and two odds
ratios are computed for each of the other levels versus that reference
category.  

But I wonder if the real problem is that serum level is a continuous
variable, and there are cut-off values corresponding to those labels.
Please clarify.  (If that is the case, things are a bit more difficult, but
not impossible. I'll wait until you clarify the situation before saying
more.)    

Cheers,
Bruce



meatduck wrote

> Hi,
>
> I'm a newbie when it comes to SPSS and Statistics, and in need of some
> help.
>
> I would like to know how to calculate a multi level factor odds ratio in
> SPSS, for an outcome (0 vs 1 : no disease vs disease) according to X
> protein
> serum levels (normal, borderline, high). And also if there's anyway to
> compare these odds ratio to another Y and Z protein.
>
> Thanks in advance,
> Carlos
>
>
>
> --
> Sent from: http://spssx-discussion.1045642.n5.nabble.com/
>
> =====================
> To manage your subscription to SPSSX-L, send a message to

> LISTSERV@.UGA

>  (not to SPSSX-L), with no body text except the
> command. To leave the list, send the command
> SIGNOFF SPSSX-L
> For a list of commands to manage subscriptions, send the command
> INFO REFCARD





-----
--
Bruce Weaver
[hidden email]
http://sites.google.com/a/lakeheadu.ca/bweaver/

"When all else fails, RTFM."

NOTE: My Hotmail account is not monitored regularly.
To send me an e-mail, please use the address shown above.

--
Sent from: http://spssx-discussion.1045642.n5.nabble.com/

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

[Nkem Ntonghanwah]

Part 1 of your problem

Logistic regression is available is SPSS via the Regression Model.

You select binary logistic 

and complete the dialogue boxes.

below two set of menus

For question 2, it would be useful to know type of design and why you wish to compare model with different dependent.

**Default model with a categorical variable 

LOGISTIC REGRESSION VARIABLES Y

  /METHOD=ENTER X 

  /CONTRAST (Agegrp_Marriage )=Indicator

  /PRINT=CI(95)

  /CRITERIA=PIN(0.05) POUT(0.10) ITERATE(20) CUT(0.5).

****Model with Reference Category for X1 is the first category, and for X2 is last  - i.e. the default

*****Variable addition using forward conditional likelihood

** summary results produced at the last step, 

LOGISTIC REGRESSION VARIABLES Y

  /METHOD=FSTEP(COND) X1 X2 

  /CONTRAST (X1)=Indicator(1)

  /CONTRAST (X2)=Indicator

  /PRINT=SUMMARY CI(95)

  /CRITERIA=PIN(0.05) POUT(0.10) ITERATE(20) CUT(0.5).

best wishes

On Thu, Jun 18, 2020 at 11:21 AM meatduck <[hidden email]> wrote:

Hi,

I'm a complete newbie when it comes to this, and in need of some help.

I would like to know how to calculate odds ratio for an outcome (0 vs 1 - no
disease vs disease), of different serum X protein levels (normal,
borderline, high) in SPSS. And is there any way to compare these odds ratio
to another Y protein or Z, with the same levels (normal, borderline, high)?

Thanks in advance,
Carlos



--
Sent from: http://spssx-discussion.1045642.n5.nabble.com/

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD